Commit 198a93b9 authored by nchernia's avatar nchernia
Browse files

Added in -v nowobble=1 for split_rmdups. This is the only non-whitespace...

Added in -v nowobble=1 for split_rmdups. This is the only non-whitespace difference in split_rmdups.awk. Dups.awk now the same as rest of Juicer, with additional code block to exit out of loop when second read end gets too far away.
parent 652760d6
Loading
Loading
Loading
Loading
+15 −0
Changes for AWS/scripts/dups.awk: 15 added lines, 0 removed lines.
Original line number Diff line number Diff line
@@ -50,8 +50,14 @@ function optcheck(tile1,tile2,x1,x2,y1,y2) {
# Executed once, before beginning the file read
BEGIN {
	i=0;
  if (length(nowobble)==0) {
      wobble1=4;
      wobble2=4;
  }
  else {
      wobble1=0;
      wobble2=0;
  }
	# names of output files
	# the variable "name" can be set via the -v flag
	dupname=name"dups.txt";
@@ -95,6 +101,10 @@ $1 != p1 || $2 != p2 || $4 != p4 || $5 != p5 || $6 != p6 || $8 != p8 || abs($3-p
						}
						else dups[k]++; #places a 1 at dups[k]
					}
					if (abs(pos1[j]-pos1[k])>wobble1) {
						#print "test", pos1[j], pos1[k]
						break
					}
				}
			}
		}
@@ -173,6 +183,11 @@ END {
							dups[k]++;
						}
					}
					if (abs(pos1[j]-pos1[k])>wobble1) {
						#print "test", pos1[j], pos1[k]
						break
					
					}
				}
			}
		}
+11 −3
Changes for AWS/scripts/split_rmdups.awk: 11 added lines, 3 removed lines.
Original line number Diff line number Diff line
@@ -26,7 +26,6 @@
# Dedup script that submits deduping jobs after splitting at known 
# non-duplicates
# Juicer version 1.5

BEGIN{
  tot=0;
  name=0;
@@ -36,7 +35,13 @@ BEGIN{
  if (tot >= 1000000) {
    if (p1 != $1 || p2 != $2 || p4 != $4 || p5 != $5 || p8 != $8) {
      sname=sprintf("%s_msplit%04d_", groupname, name);
      if (justexact) {
	sysstring = sprintf("bsub -W %s -o %s -q %s -J %s <<- EOF\nawk -v nowobble=1 -f  %s/scripts/dups.awk -v name=%s/%s %s/split%04d;\nEOF\n", queuetime, outfile, queue, sname, juicedir, dir, sname, dir, name, dir, name);
      }
      else {
	sysstring = sprintf("bsub -W %s -o %s -q %s -J %s <<- EOF\nawk -f  %s/scripts/dups.awk -v name=%s/%s %s/split%04d;\nEOF\n", queuetime, outfile, queue, sname, juicedir, dir, sname, dir, name, dir, name);
      }
      
      system(sysstring);
      if (name==0) {
	waitstring=sprintf("done(%s)", sname);
@@ -55,7 +60,12 @@ BEGIN{
}
END {
  sname=sprintf("%s_msplit%04d_", groupname, name);
  if (justexact) {
    sysstring = sprintf("bsub -W %s -o %s -q %s -J %s <<-EOF\nawk -v nowobble=1 -f %s/scripts/dups.awk -v name=%s/%s %s/split%04d;\nEOF\n", queuetime, outfile, queue, sname, juicedir, dir, sname, dir, name, dir, name);
  }
  else {
    sysstring = sprintf("bsub -W %s -o %s -q %s -J %s <<-EOF\nawk -f %s/scripts/dups.awk -v name=%s/%s %s/split%04d;\nEOF\n", queuetime, outfile, queue, sname, juicedir, dir, sname, dir, name, dir, name);
  }
  system(sysstring);
  if (name==0) {
    waitstring=sprintf("done(%s)", sname);
@@ -67,6 +77,4 @@ END {
  system(sysstring);
  sysstring = sprintf("bsub -W %s -o %s -q %s -J %s_rmsplit  -w \"done(%s_catsplit)\" <<- EOF\n rm %s/*_msplit*_optdups.txt; rm %s/*_msplit*_dups.txt; rm %s/*_msplit*_merged_nodups.txt; rm %s/split*;\nEOF",queuetime, outfile, queue, groupname, groupname, dir, dir, dir, dir);
  system(sysstring);


}